PaliGemma 2
Open vision-language model family in 3B, 10B, and 28B sizes.
About
PaliGemma 2 is Google's open vision-language model family pairing the SigLIP-So400m vision encoder with Gemma 2 language models in 3B, 10B, and 28B sizes, at 224px, 448px, and 896px resolutions. It improves on the original PaliGemma with richer image descriptions — actions, emotions, and scene context — plus state-of-the-art results on OCR-heavy tasks like table and music-score recognition after fine-tuning.
Like all VLMs it can hallucinate, and the base models are meant as fine-tuning starting points rather than chatbots.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
SOTA on OCR transfer tasks (table/molecule/music-score recognition) after fine-tuning (Google-reported).
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: broad Google VLM family; 3 sizes x 3 resolutions; strong OCR fine-tuning results; Gemma license terms apply; base models are fine-tuning starting points, not chatbots. License: Gemma Terms of Use (custom Google open-model license)..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 8.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · broad Google VLM family; 3 sizes x 3 resolutions; strong OCR fine-tuning results
- · Gemma license terms apply
- · base models are fine-tuning starting points, not chatbots